close

Вход

Забыли?

вход по аккаунту

?

Laser photodissociation of organometallic compounds on a cryosubstrate.

код для вставкиСкачать
APPLIED ORGANOMETALLIC CHEMISTRY, VOL. 5,247-255 (1991)
Laser photodissociation of organometallic
compounds on a cryosubstrate
Masahiro Kawasaki* and Nobuyuki Nishit
*Research Institute of Applied Electricity, Hokkaido University, N12W6, Sapporo 060, Japan, and
?Institute for Molecular Science, Myodaiji, Okazaki 444, Japan
The photodissociation dynamics of organometallic
compounds (tetramethyltin, trimethylgallium,
trimethylindium and dimethylzinc) adsorbed on a
quartz substrate at 100K have been studied by
time-of-flight mass spectrometry, detecting mainly
CH; photofragments. Shapes of the time-of-flight
spectra depend on the flux of the dissociation laser
at 193nm and the thickness of multilayers of
organometallic compounds. In thick layers, not
only a low energy component but also a high
energy component are observed in the time-offlight spectrum of the CH3photofragments. In thin
layers, relaxation processes take place so quickly
that the product time-of-flight profiles are characterized by a Maxwell-Boltzmann temperature
profile.
Keywords: Photodissociation,
organometallic,
time-of-flight, translational energy, dynamics,
photoejection, excimer laser
1 INTRODUCTION
Laser surface chemistry has been used as a basis
for many new techniques in surface processing,
including photochemical vapor deposition of
metals, semiconductors and insulators, which are
potentially useful for the microelectronics
industry.14 Analytical efforts to study the laser
photolysis of adsorbate systems are only just
beginning: spectroscopic techniques used recently
for surface photochemistry are time-of-flight
(TOF) mass spectroscop >11 time-resolved laserinduced fluorescence, 1 2 , l Fin situ ESCA14and UV
spectro~copy.'~
The irradiation by photons of organic and inorganic compounds adsorbed on substrates leads to
a nonthermal photochemistry that is different
from that observed in the gas phase.
Photodissociation of monolayers on substrates
has been reported by some groups. They have
measured the translational energy distribution of
0268-2605/91/040247-09$05.OO
01991 by John Wiley & Sons, Ltd.
photofragments from monolayers of chlorine CI2,'
HzS,697 CH3C1,' CH3Br,6*7trimethylaluminum
A1(CH3),10 dimethylzinc Zn(CH3)2,11trimethylgallium (Ga(CH,),)
and trimethylindium
(In(CH,),)"~ l3 ejected from the substrates. The
dynamics of the surface photodissociation processes were found to vary with the coverage of the
adsorbed layer^."^ In general, for higher coverage, both slow and fast photofragments are
ejected from the substrate by laser irradiation
whilst for lower coverage only slow photofragments are observed."
In the photodissociation of trimethylgallium on
a substrate at room temperature, gallium atoms
are detected by laser-induced fluorescencel2.l3
whilst aluminum atoms are not detected by a
time-of-flight technique." Higashi" has stated
that the absence of aluminum signal is particularly
significant because this substantiates the view that
aluminum atoms are strongly bound to the surface. In the gas phase, organometallic com ounds
strongly absorb UV photons at <200 nm,"resulting in the formation of methyl radicals and other
fragments. For example, the photodissociation of
gaseous tetramethyltin at 193 nm has been
reported. l7 Photoexcited molecules dissociate into
both Sn(CH,), + CH3 and Sn(CH,), + 2CH; by
two-body and three-body dissociation processes,
respectively. Cd(CH,), and Zn(CH3)2form electronically excited CH3Cd' and CH,Zn. radicals by
193 nm irradiation in the gas phase.'' In order to
understand how the nature of photon-adsorbate
interactions change in different environments,
we investigated the UV photodissociation of
monolayers and multilayers of molecules on a
cryosubstrate.
2
EXPERIMENTAL
The experimental apparatus'' is drawn schematically in Fig. 1. The reaction and detector
chambers are differentially pumped by turbomolecular pumps and a diffusion pump. The pressures of these chambers were
and lo-'' Torr.
Received 18 February 1991
Accepted 13 April I991
M KAWASAKI AND N NISHI
248
F
\\ +
J/
a
<+$&ubstrate
I/
Laser
b
-- .__
Figure 1 Schematic diagram of the experimental set-up for
solid photolysis. The hollow arrow indicates the laser beam.''
respectively. The quartz substrate is cooled to
100K by a liquid nitrogen trap, on which the
sample molecules, dimethylzinc (Zn(CH3)2),trimethylindium (In( CH,),),
trimethylgallium
(Ga(CH,),) or tetramethyltin (Sn(CH,),), are
deposited through a capillary tube until they form
multilayers on the substrate. An ArF excimer
laser (193 nm) is used to irradiate the molecules
on the substrate at glancing incidence (800 to the
normal) with unpolarized radiation. The laser
power used is 50-100 mJ cm-'.
Radicals ejected from the substrate were
detected by a quadrupole mass spectrometer with
an electron-bombardment ionizer under conditions of electron energy = 44 V for the Sn(CH3),
experiment and 20 eV for other compounds, ion
energy = 20 V as a function of mass numbers and
time after laser pulses. The flight length was
16 cm.
High-purity sample gases Zn(CH3)2, In(CH3),
and Ga(CH3)3 were supplied by Sumitomo
Chemical Co. Research-grade Sn(CH3), was purchased from Wako Chemicals.
3 RESULTS
L
0
100
Time of
200
flight
300
/)JS
Figure 2 Time-of-flight spectra of CH3 radicals obtained in
Sn(CH,), photolysis at 193nm on the quartz substrate cooled
at l00K and in the gas phase. (a) Low laser energy flux
(50 mJ cm-'); (b) high laser energy flux (100 mJ crn-'); (c)
gas-phase photolysis.
peak TOF signals obtained from these figures are
summarized in Table 1.
A comparison of the time-of-flight profile of the
CH: signal (Fig. 2) with that of the Sn(CH3):
signal (Fig. 3) reveals that the CH: ion does
originate neither from the dissociative ionization
of parent molecules nor from tin-containing
photofragments, but rather by low-energy (44 V)
electron bombardment ionization of CH; photofragments. This was also reported in the multiphotoionization and low-energy (30 V) electron
bombardment of CH; fragments produced from
VI
c
.-
3.1 Photodissociation of
tetramethyltin at 193 mm
3.1.1 Time-of-flight distribution of photofragments
The laser-irradiated Sn(CH3)4 molecules were
deposited on the quartz plate cooled to 100K.
Photofragments were detected by the mass
spectrometer with an electron bombardment
ionizer. Detected ions were CH: and Sn(CH,):
(m= 0-4). Some of these ions were generated in
the ionizer by dissociative ionization of neutral
photofragments or parent molecules desorbed.
The TOF spectra of these species were measured
as shown in Figs 2 and 3. Both the threshold and
0
800
1600 0
Time of flight
800
1600
/ ps
Figure 3 Time-of-flight spectra of Sn(CH,); obtained in
photolysis of solid Sn(CH3), at 193nm on the quartz substrate
cooled at 1 0 0 K . Detected ions are (a) Sn+; (b) Sn(CH,):;
(c)SnCH: ; (d) Sn(CH3)$.
249
LASER PHOTODISSOCIATION ON A CRYOSUBSTRATE
Table 1 Time-of-flight results and average energies of photofragments from the
photodissocaitionof organornetalliccompounds at 193 nm on cryosubstrate at 100 K
Parent
Molecule
Detected
Species (amu)
tlHa
(ps)
(I4
Sn(CH3)4
CH; (15)
Sn' (120)
SnCH: (135)
Sn(CH3): (165)
Sn(CH,): (180)
CH: (15)
CH: (15)
CH: (15)
CH: (15)d
23+2
150+5
114f 10
120+_ 10
140 f 5
25+2
29+2
100
33/68
330
250
250
310
701270
60/130/220
60/210
-180
In(CH,),
Ga(CH3),
Zn(CH3),
-
tpb
ETTIC
ET;
kJ mol- ')
2.1
1.7
7.1
2.1
0.8
2.9
3.4
7.1
42
21
37.8
29.4
25.2
37.8
21
-
Threshold appearance time-of-flight. Peaked time-of-flight. ' Assuming bimodal
Maxwell-Boltzmann distributions of Eqn [6], I ? , = (3/2)k,. Experiment on a room
temperature substrate.
a
the photodissociation of solid Zn(CH,), by
Howitz et al." Figure 2(a) and (b) shows the laser
power dependence of the TOF shapes of the CH;
photofragments. For Fig. 2(b) the laser energy
flux is twice as large as for Fig. 2(a). At the high
laser energy flux the TOF spectrum shifts toward
the faster region, peaking at 70ps. At low flux it
peaks at 130ps. For purposes of comparison, the
result of the gas-phase photodissociation of a
Sn(CH3)4 molecular beam is shown in Fig. 2(c),
which was measured previously with the same
spectrometer that was used for this study.17Figure
2(c) shows that the TOF of CH; peaks at 80ps in
the molecular beam experiment, which is close to
that obtained in Fig. 2(b). At high energy flux,
parent molecules may be desorbed from the solid
by laser irradiation, resulting in gas-phase photodissociation during the laser pulse duration of
15 ns. viz.
Hencz, the strong and slow peak observed at
TOF= 130ps in Fig. 2(a) is not due to gas-phase
photodissociation but to photodissociation of
solid Sn(CH3)4.
The fast weak peak of Figs 2(a) and (b) starts to
appear at 23ps and shows a hump at 33ps. If
these signals are due to the solid-phase photodissociation of Sn(CH3)4,the fastest CH; radicals
can carry all the excess energy, E,, = hv - D o , as
translational energy because the substrate carries
no translational energy. The threshold TOF of
23 k 2 p s corresponds to 360 k kJ mol-', in good
agreement with that calculated with hv =
619 kJ mol-l and D,,
= 250 k 10 kJ mol-I.
We must consider the fact that Sn(CH,)L
(m= 0-3) can be generated by dissociative ionization processes in the ionizer. The parent molecule
is promptly photodesorbed by UV laser irradiation. The strong Sn+ signals are produced by
dissociative ionization of Sn(CH3)4because (i) the
TOF shape of Sn' in Fig. 3(a) resembles that of
Sn(CH3)4+in Fig. 3(b) and (ii) the threshold TOF
of Sn+(150+5ps) is in agreement with that of
Sn(CH4): (140 k 5 p s ) . Similarly, the most probable source of SnCH: is dissociative ionization of
Sn(CH,):, because (i) the shape of Fig. 3(c)
resembles that of Fig. 3(d) and (ii) the threshold
TOF of 114k lops for SnCH: is in agreement
with 120 k lops for Sn(CH,):. However, when
one compares carefully the TOF spectra of
Sn(CH3): with the spectrum of SnCH; one finds
that the Sn(CH,): signals are stronger in intensity
at TOF>600ps than the SnCH: signal. Part of
this Sn(CH3): signal must be generated by the
dissociative ionization of the parent molecules
that are desorbed by laser irradiation. Hence, the
true TOF spectrum for Sn(CH,); neutral photofragments is represented by that of SnCH: and
not by that of Sn(CH,):.
Translational energy distribution of
photofragments
The TOF spectra of the photofragments are converted to a translational energy distribution P(&)
with a suitable Jacobian f a c t ~ r . For
' ~ the CH;
photofragments, the energy distribution is
obtained at a laser flux of 50 mJ cm-' (Fig. 4(a)),
3.1.2.
M KAWASAKI AND N NISHl
250
'
1
-
shown in Fig. 5(b). There seem to be two energy
distributions; one peaks near zero and the other
at 10 kJ mol-'. The average translational energies will be discussed below.
a
I
-
I
3.2 Photodissociation of
trimethylindium at 193 nm
0 5
Figure 4 Translational energy distributions, P(E,), of CH;
radicals obtained in solid and gas photodissociation of
Sn(CH3)4at 193 nm. Dotted data are from Fig. 2. Solid lines
are simulated curves from Eqn [6] with parameters from Table
1. (a) Low laser energy flux (501nJcm-~); (b) high laser
energy flux (lo0mJ cm-*; (c) gas-phase photolysis.
100mJcm-' (Fig. 4(b)), or for the gas phase
photodissociation (Fig. 4(c)). At 50 mJ cm-' the
translational energy distribution peaks at
8.4 kJmol-' while at 100mJcm-2 and for the
gas-phase photodissociation, the distribution of
CH3 peaks at 17 kJ mol-'. As mentioned above,
at high laser energy flux, the parent Sn(CH3)4
molecules photodesorbed by laser irradiation
undergo photodissociation near the substrate.
The energy distribution of Fig. 4(b) which peaks
at 17 kJ mol-' resembles that of the gas-phase
photolysis of Fig. 4(c). Hence, the distribution of
Fig. 4(a), i.e. the low-flux case, shows the nasceni
distribution of CH; radicals produced from the
photodissociation of solid Sn(CH3)4. In the solidphase dissociation, P(&) peaks at low kinetic
energy. This component is characteristic of a fast
relaxation process in the solid photolysis.
The TOF signals of the parent ions Sn(CH3):
are converted to the translational energy distribution of photodesorbed Sn(CH& as shown in Fig.
5(a). However, a part of the Sn(CH,): signals
were generated by dissociative ionization of
in the ionizer; the TOF distribution of
Sn(CH3): is not analyzed. Instead, the SnCH:
signal was converted to P(&) as the translational
energy distribution of Sn(CH3;. The result is
In(CH3)3 deposited on a quartz substrate cooled
to 100 K was irradiated with 193 nm laser light.
CH: ions were detected by the mass spectrometer
as shown in Fig. 6. As stated above, CH: signals
originate from CH; photofragments under the
experimental conditions of low electron energy
for ionization (20 V). Hence, Fig. 6 gives the TOF
spectrum of the CH; photofragments from solid
In(CH3), . Both a fast and a slow signal peaks are
observed. The threshold TOF is observed at 25 k
2 ,us and the peak TOF signal is at 70 k 5 ,us for the
fast component and at 270 k 5 ,us for the slow one.
This bimodal distribution is typical in photodissociation of solid molecules, as is observed in
Fig. 2(a). The TOF spectrum of this CH; radical is
converted to the translational energy distribution
as given in Fig. 7. One peaks at low translational
energy and the other at
10 kJ mol-'. Average
translational energies are 3 and 38 kJ mol-*
assuming a Maxwell distribution for the velocities.
-
'y
7
i1
c
ul
05
4.;-
Translational Energy/
I
kJ rnol.'
Figure 5 Translational energy distributions P(ET) of
Sn(CH,), parent molecules (a) and Sn(CH,)j radicals (b)
obtained in laser irradiation of solid Sn(CH,), at 193 nm.
Sn(CH3)4signals are observed as Sn(CH,): whilst Sn(CH);
signals are observed as SnCH: by dissociative ionization in the
ionizer. Solid lines are simulated P(ET) from Eqn [6] with
parameters from Table 1.
25 1
LASER PHOTODISSOCIATION ON A CRYOSUBSTRATE
f
3
/
4 7681
-z
0)
.m
1
.
0-4
o
LOO
2000
800
Time of flight
/
ps
Figure 6 Time-of-flight spectrum of CH3 radicals obtained in
solid In(CH3); photodissociation at 193 nm on the quartz
substrate cooled at 100 K. The detected species is CH; .
3.3 Photodissociation of
trimethylgallium at 193 nm
The TOF spectrum of CH: signals from photodissociation of solid Ga(CH3), (Fig. 8) shows a
rather complicated shape. TOF spectrum peaks at
60ps, 130ps, and 220ps are marked by arrows in
Fig. 8. This may be compared with the photodissociation of Sn(CH3)4and In(CH3)2which have
only two peaks (Figs 2 and 6). This fact may
reflect a rather complicated dissociation process
for solid Ga(CH3)3.
3.4 Photodissociation of dimethylzinc
at 193 nm
Zn(CHJ2 deposited on the quartz substrate
cooled to 100 K was irradiated with a laser beam.
The CH; signals show a bimodal distribution, as
in Fig. 9(a). The threshold TOF is 29 k 2ps. TOF
Translational
Energy
/
Time of flight/Ps
Figure 8 TOF spectrum of CH; radicals obtained in photolysis of solid Ga(CH3)3at 193 nm on the quartz substrate cooled
at 100K. The arrows show the positions of peaks in the
spectrum.
peaks are at 60 and 210ps as summarized in Table
1. When the substrate was warmed up to room
temperature, Zn(CH3)* was desorbed, leaving a
monolayer of Zn(CH,), on the substrate. Laser
irradiation of this substrate gave CH3 signals as
shown in Fig. 9(b), peaking at 18Ops. This peak
position is close to the slow one observed at
210ps in the photodissociation of solid Zn(CH,),
shown in Fig. 9(a).
These TOF spectra are converted to P(&) as
shown in Fig. 10. There seem to be two different
energy distributions in Fig. lO(a) for the solid
photodissociation, while only the low-energy
component appears in Fig. 10(b) for dissociation
of the monolayer.
0
kJ mol-’
Figure 7 Translational energy distribution of CH; radicals
obtained in photodissociation of solid In(CH3), at 193 nm.
The solid lines are simulated curves from Eqn [6] with parameters from Table 1.
400
Time of flight
800
12 I0
/ ps
Figure9 TOF spectrum of CH; radicals obtained in
Zn(CH& photodissociation at 193nm (a) on a quartz substrate cooled at 100 K, and (b) on a quartz substrate at 300 K.
M KAWASAKI AND N NISHI
252
of Fig. lO(a) is asttributable to the direct dissociation of the topmost layer of Zn(CH3),, viz.
+
CH3ZnCH3/multilayer hv-+CH; (fast) [3]
For the other compounds, Sn(CH3)4 and
In(CH3)3, this direct photodissociation is
observed as shown in the high-energy components of Figs 2(a) and 6. In the direct photodissociation process, the fastest CH3 radicals can
carry all the excess energy as the translational
energy. The maximum translational energy is
given by Eqn [4]:
a
0
10
Translational
20
Energy
J
30
40
/
kJ mo1-l
Figure 10 Translational energy distribution of CH3 radicals
obtained in photodissociation of solid Zn(CH3)* at 193nm.
The solid lines are simulated curves from Eqn 161 with parameters from 'Table 1 . (a) Solid photodissociation on a quartz
substrate cooled to 100 K; (b) nionolayer photodissociation on
a quartz substrate at room temperature.
4
DISCUSSION
4.1 Effect of thickness of multilayers
of organometallic compounds on TOF
distributions
As mentioned in Section 3.4 above, the photo-
dissociation of Zn( CH3)?molecules monolayered
on a quartz substrate at room temperature gives
only the slowly moving photofragments or the
low-energy component signals, whilst photodissociation of the multilayered molecules gives
not only a low-energy component but also a highenergy one. A similar dependence of kinetic energies on the surface coverage has been reported for
the photodissociation of chlorine and CH3C1 on
silicon wafers.' The TOF distributions of chlorine
photofragments are contrasted for thin and thick
depositions of parent C1, molecules on the substrate. The TOF distribution is bimodal for the
thin-deposition case whilst it is unimodal and
contains only a high-energy component for the
thick-deposition case. The high-energy component of the photofragments originates from the
topmost layer of multilayered Clz and CH3Cl
molecules; hence the position of threshold
appearance time is shifted toward longer time as
the photon energy decreases. This is also the case
for solid photodissociation of CH212on A1203and
aluminum surface^.^ In the present experiment;
the high-energy component in the TOF spectrum
ET(max)= hv - Do- AE,,
[41
where AE,,, is the difference in adsorption energies of a parent molecule and of the CH3 radicals
from the solid surface, and Do is the bond dissociation energy which is tabulated in Table 2. The
adsorption energy is a function of adsorption
geometry with atom(s) bonded to the solid surface and is typically a few kilojoules per mole. For
the photodissociation of solid Sn(CH3)4, the
threshold TOF corresponds to 360 f!:T kJ mol-'.
This value gives AE,,,=O using Eqn [4] with
hv = 619 kJ and Do= 250 kJ mol-'. AEads is
expected to be close to zero in this case since
S ~ I ( C € Tis~ adsorbed
)~
by the CH3 attached to tin
and the leaving species is a CH; radical. For
In(CH3), AE,,= 110?:! kJ mol-' is obtained from
the observed ET(,,,=) of 310:; kJ mol-' and Do=
200 kJ mol-'. AEads could be larger in this case
because In(CH3), is adsorbed by interaction
between the solid surface and the indium metal
while the leaving CH; radical can interact with the
surface through its carbon or hydrogen atoms.
For Zn(CH,), ,the observed maximum translational energy of the CH3 photofragment is
230'g kJ mo1-', which is much lower than that of
Table 2 Thermodynamics of the decomposition of organometallic compounds
Reaction
+ CH,
Sn(CH,)? +2CH,
Sn
+4CH,
Ga(CH,),+ Ga(CH1)Z+ CH,
In(CH,), +In(CH& + CH,
Zn(CH3)*-*ZnCH3 + CH,
Sn(CH&
-+Sn(CH,),
AH(kJ mol-')
Ref
25 I
498
907
251
23.24
23,24
23,24
25
25
26
197
268
LASER PHOTODISSOCIATION ON A CRYOSUBSTRATE
the other organometallics studied. This is presumably due to the large difference in adsorption
energies, AEads,between Zn(CH3)2and CH;.
Tabares et al.’ have observed a bimodal energy
distribution in the photofragmentation processes
of a multilayer of CH3Br at 193nm. They attribute this slow and wide distribution to the collision of CH; radicals within the deposit before the
CH; fragments can emerge from the surface. This
is possible because the optical depth of CH3Br at
193 nm is 1200 nm. This seems, however, not to
be the case in the present experiment because
multilayered Zn(CH& gives signals of both highand low-energy components but monolayered
Zn(CH3)2 yields only a low- and wide-energy
component.
The
low-energy component
appeared in the thin-deposition case of chlorine
and CH3CI photodissociation.’ This result suggests that the strong interaction between desorbates and substrate decelerates fragment velocity
and also makes its distribution wide. Higashi” has
stated in his report on photodissociation of monolayered Al(CH3)3that the dissociative transition is
not direct since so little energy is imparted to the
ejected methyl radicals. This is consistent with
the fact that subthermal desorption distributions
have been observed in thermal desorption
experiments.
Chuang and Domen’ have explained that, in
thick molecular layers, many of the photoexcited
molecules beneath the surface may not dissociate
because of repulsive interaction with the surrounding molecules as the molecular bond distance expands in these surrounding molecules. In
addition, Bourdon et aL6 have reported in their
photodissociation of CH3Br on a LiF substrate
that the mean translational energy distribution of
the photofragments is independent of the detction
angle and, therefore, independent of the depth of
the desorbing gas that is being traversed. This
indicates that collisions between photofragments
and desorbate are not important in the determination of the translational energy distribution.
4.2 Translational energy distributions
of photofragments from solid phase
photodissociation
The TOF spectra of the CH; photofragments
from the organometallic compounds have rather
wide peaks in the present experiment. The
FWHM is as wide as 1OOps. This wide distribution reflects quick energy flow among the oscillators of solid molecules and the lattice mode of
253
the substrate. Lin and his co-workers20have developed a theoretical model for the photodesorption
using transition state theory. The energy of the
activated complex is divided into two parts: vibrational and translational energies. The calculation
of distribution in the translational energy ET is
essentially a problem of calculating the distribution of the vibrational energy among a collection
of oscillators in the multilayered molecules. The
vibrational energy of the activated complex along
the reaction coordinate becomes a part of the
translational energy of the dissociation products.
In this case the translational energy distribution is
given by Eqn [5]:
Hence, T is a characteristic temperature of this
distribution. We assume this equation is applicable to the photodissociation reaction on the
substrate and in the solid phase. Figures 4, 5, 7
and 10 depict the translational energy distributions, P(ET), of CH3 photofragments obtained
from corresponding TOF spectra along with bestfit curves. Since two primary processes occur in
the photodissociation as described above, a twocomponent fit is tested. The smooth curves
through the experimental traces are fitted to composites of two Maxwell-Boltzmann distributions
as given by Eqn [6]:
Calculated curves are fitted fairly well to the
experimental results. Table 1 summarizes the
results of average energies defined as ETi=
(3/2)kTi.
Concerning the results of Sn(CH3),, when one
compares Fig. 4(b) with (c), the energy distribution above 8 kJ mol-’ in Fig. 3(b) is similar to that
of Fig. 4(c), suggesting that at high laser energy
flux, gas-phase photodissociation occurs. The
low-energy components (average energy
2 kJ mol-’ ) reflect surface photodissociation. This
subthermal desorption is attributable to shallow
potential wells for desorption.21 The concept of
what is occurring can be understood by performing an imaginary experiment of scattering molecules from surfaces and invoking microscopic
reversibility. For shallow potential wells, highenergy incoming molecules can be imagined to
254
scatter with zero residence time. Low-energy
molecules which scatter inelastically are easily
trapped. In equilibrium, the flux-out is equal to
the flux-in and therefore leads to a Maxwellian
distribution. In photodesorption, however, only
the molecules resident on the surface from prior
trapping events contribute to the distribution. In
this case the distribution cannot look Maxwellian
if microscopic reversibility is to be satisfied. This
concept explains the fact that Figs 4,7 and 10 for
CH; radicals from SII(CH,)~, In(CH,), and
Zn(CH3)*do not completely fit to a Maxwellian
distribution.
The parent molecule of Sn(CH3), is photodesorbed with comparatively large translational
energy (-25 kJ mol-') as shown in Table 1. In the
case of chlorine, it is rather small (-4 kJ mol-').'
This difference may be attributable to the different nature of electronic states excited by laser
irradiation. A Rydberg state can be excited for
SII(CH,)~by 193 nm irradiation, whilst a valence
state for chlorine. Repulsion forces between
Rydberg orbitals ejects Sn(CH3), molecules from
the surface because of the large diameters of
Rydberg orbitals.
M. KAWASAKI AND N . NISHI
LIF signals of zinc atoms.', It can be estimated
from the Boltzmann energy distribution of the
photofragment translational energy that the
energy dissipation rate on a substrate is large. The
atoms of the low-temperature-melting-point
metals gallium and indium are photoejected,
whilst the high-temperature-melting-point metals
aluminum and zinc are not. By experimentally
determining if metal atoms are photoejected, the
nature of the bond between newly formed metal
atoms and a substrate can be investigated.
4.4 Photodissociationof
trimethylgallium
When Ga(CH,), is photodissociated in the solid
phase (Fig. 8) CH; photofragments show a complicated TOF spectrum. This result suggests that
the CH; radicals as generated from at least two
processes, i.e. one from the parent molecule,
Ga(€H3)3, and the other from a metastable molecular fragment, probably GaCH;. Mitchell et ~ 1 . ' ~
proposed that gas-phase photodissociation of
Ga(CH,), at 222 nm produces GaCHj. Since this
species absorbs another UV photon, GaCH; photolysis acts as a 'bottleneck' in the production of a
free gallium atom and also a CH; fragment. It is
4.3 Photoejection of metals atoms
interesting to note that a similar situation pertains
In the solid-phase photodissociation of SKI(CH~)~, in the pyrolysis of Ga(CH3)3r where under
certain conditions GaCHj remains as a polymer
Sn' signals are detected but these signals origideposit after decomposition of Ga(CH3)3 and
nate from dissociative ionization of photodeGa(CH3);.I4
sorbed parent molecules and not from photofragment tin metals, as described in Section 3.1.1
Acknowledgement M Kawasaki thanks the Cooperative
above. In fact, Table 1 shows that Sn+ and
Study Program of the Institute for Molecular Science. This
Sn(CH,): have almost the same TOF thresholds
work is partly supported by a Grant-in-Aid on Priority Area
(tm) and average energies (ET).Tin metal atoms
Research on 'Photo-excited process' supported by the
are not photodesorbed in our experimental conMinistry of Education, Science, and Culture of Japan.
ditions. Based on the AH values shown in Table
2, the formation of tin metal requires at least two
photons at 193 nm and this process is reported to
REFERENCES
play a minor role compared with a one-photon
pro~ess.'~
Higashi" has reported the absence of
aluminum in his surface photodissociation of
1. Osgood, R M and Deutsch, T F Science, 1985, 227: 709
2. Yardley, J T Laser Handbook Bass, M and Stitch, M L
A1(CH3)3, suggesting that aluminum is strongly
(eds) Elsevier Science Publishers, Netherlands 1985
bound to the surface. The two-photon photolysis
3. Hanabusa, M Mat. Sci. Rep., 1987, 2: 51
of Ga(CH,), adsorbed on a quartz plate has been
4 . Herman, I P Chem. Reo., 1989, 89: 1323
studied at 243nm by probing gallium atoms by
5. Kawai, T and Sakata, T Chem. Phys. Lett., 1980, 69: 33
laser-induced fluorescence (LIF) .l2 When a gal6. Bourdon, E B D, Cowin, J P, Harrison, I , Polanyi, J C,
lium metal plate was irradiated by the same laser
Segner, J, Stanners, C D and Young, R A J. Phys. Chem.,
beam, no LIF signals were detected. This is also
1984, 88: 6100
the case for In(CH,), .13 Thus, gallium and indium
7. Tabares, F L, Marsh, E P, Bach, G A and Cowin, J P J .
atoms originate from adsorbed Ga(CH,), and
Chem. Phys., 1987,86: 738
III(CH~)~,
respectively. Photolysis of adsorbed
8. Kawasaki, M, Sato, H and Nishi, N J . Appl. Phys., 1988,
65: 792
Zn(CH,)* on a quartz substrate gave, however, no
LASER PHOTODISSOCIATION O N A CRYOSUBSTRATE
9. Chuang, T J and Domen, K J. Vac. Sci. Tech., 1987, A5:
10.
11.
12.
13.
14.
15.
16.
17.
473
Higashi, G S J. Chem. Phys., 1988, 88: 422
Howitz, J S, Villa, E and Hsu, D S Y J. Phys. Chem.,
1990, 94: 7214
Suzuki, H , Mori, K, Kasatani, K, Kawasaki, M and Sato,
H J. Appl. Phys., 1988, 64: 371
Kawasaki, M, Kasatani, K, Sato, A and Nishi, N Mat.
Res. SOC. Sym. Proc., 1989, 129: 69
Shogen, S, Matsumi, Y, Kawasaki, M and Okabe, H J.
Appf. Phys. (in press)
Shaw, P S, Sanchez, E, Wu, Z and Osgood, R M Jr Chem.
Phys. Lett., 1988, 151: 449
Thompson, H W J. Chem. SOC., 1936: 108
Kawasaki, M and Sato, H Laser Chem., 1987, 7: 109
255
18. Yu, C F, Youngs, F, Ysukiyama, K and Bersohn, R J.
Chem. Phys., 1986, 85: 1382
19. Nishi, N , Shinohara, H and Okuyama, T J . Chem. Phys.,
1984, 80: 3898
20. Lin, S H , Tsong, I S T, Ziv, A R , Szymonski, M and
Loxton, C M Physica Scripta, 1983, T6: 106
21. Tully, J C Surf. Sci., 1981, 111: 461
22. Mitchell, S A, Hackett, P A, Rayner, D M and
Humphries, M R J . Chem. Phys., 1985, 83: 5028
23. Lampe, F W and Niehaus, A J. Chem. Phys., 1968, 49:
2949
24. Gedanken, A, Robin, M B and Kuebler, N A J. Phys.
Chem., 1982, 86: 4096
25. Kerr, J A Chem. Rev., 1966, 66: 465
26. Jackson, R L Chem. Phys. Lett., 1989, 163: 315
Документ
Категория
Без категории
Просмотров
1
Размер файла
696 Кб
Теги
cryosubstrate, organometallic, compounds, photodissociation, laser
1/--страниц
Пожаловаться на содержимое документа